Empiric antibiotic coverage of atypical pathogens for community acquired pneumonia in hospitalized adults
Abstract
This is a protocol for a Cochrane Review (Intervention). The objectives are as follows:
This review aims to evaluate the efficacy and need of adding atypical coverage in patients hospitalised due to CAP. In doing so, this review intends to estimate the mortality and efficacy of regimens containing atypical antibiotic coverage.
Secondary objectives are:
-
To evaluate the frequency of adverse effects associated with different types of antibiotic treatment.
-
To evaluate the efficacy of regimens containing atypical antibiotic coverage for patients with identified atypical pathogens (see later for definition);
-
To evaluate the efficacy of regimens containing atypical antibiotic coverage for patients with identified Streptococcus pneumoniae.
Outcomes will be collected for all patients, and separately for patients with identified atypical pathogens, and those with pneumococcal infection.
Background
Community acquired pneumonia (CAP) is a common clinical entity with considerable mortality and overwhelming economical burden. The Pneumonia Patient Outcomes Research Team (PORT) has developed a prediction scheme utilizing clinical variables to stratify CAP patients into five mortality risk classes, by which 30 day mortality has been estimated up to 9.3% and 31.1% for patients stratified to risk class IV and V, respectively (Fine 1997). Though Streptococcus pneumoniae is the leading cause, frequency of other agents, especially those designated "atypical pathogens" (Legionella sp., Mycoplasma Pneumoniae and Chlamydia Pneumoniae), varies with geographical area, age groups and diagnostic means available. Distinction between typical and atypical organisms is significant, as the latter share intracellularity and are treated with specific antibiotic drugs. The causing agent cannot be clearly sorted to either group by any clinical, radiological or laboratory parameter (Donowitz 2000; Levison 2001). As the microbiological diagnosis is often unknown, initial antibiotic treatment of CAP is largely empirical.
All major guidelines for treatment of CAP divide patients to three subgroups: Outpatients, hospitalized patients on a general ward and hospitalized patients in intensive care units. Major guidelines recommend in general an antibiotic regimen covering both typical and atypical pathogens in hospitalized CAP patients (IDSA 2000; Mandell 2000; ATS 2001; BTS 2001).
Suggested regimens include extended spectrum cephalosporins/ amoxicillin / beta‐lactam/beta‐lactamase inhibitors combined with macrolides, or a single agent with broad coverage, i.e. fluoroquinolones. British Thoracic Society (BTS) guidelines support this recommendation in view of a high percentage of atypical pathogens, particularly Legionella sp., in hospitalized patients in the UK (higher, incidentally, than in North America or Europe). The American Thoracic Society (ATS) reserves recommendation of atypical coverage to clinical circumstances, though conceding elsewhere that clinical presentation cannot establish diagnosis. All guidelines support extensive antibiotic coverage for intensive care unit (ICU) patients, noting the prevalence as well as high mortality rates of pneumonia caused by Legionella sp. in that population. Admittedly, these guidelines are based on insufficient evidence, as there have been only few randomised controlled trials assessing CAP treatment protocols.
Most trials comparing different treatment arms were performed to evaluate a specific drug, frequently a macrolide or a fluoroquinolone. Several of these trials (incidentally) compared regimens with and without atypical antibiotic coverage, and found results to be generally equal (Aubier 1998; Tremolieres 1998). Furthermore, a large percentage of atypical pneumonia cases have been shown to be in fact dual infections, usually in combination with a typical pathogen (Donowitz 2000). In view of these data, one must wonder what the true role of atypical agents in pneumonia really is: If patients improve regardless of atypical coverage, should this coverage be routinely included? We have not found a single prospective randomised controlled trial comparing as an objective a treatment regimen with atypical antibiotic coverage to one without such coverage. Furthermore, no systematic review surveying trials that include such a comparison has been identified. Expanding coverage to atypical pathogens is bound to increase toxicity, resistance and cost. Moreover, this broad coverage might be at the expense of efficacy of pneumococcal coverage (Muder 1991; Leibovici 1992).
In this systematic review, concentrating on hospitalised adults with CAP, all trials comparing treatment regimens with and without atypical antibiotic coverage will be extracted, in hope to evaluate the efficacy and the need of adding atypical coverage in that population.
Objectives
This review aims to evaluate the efficacy and need of adding atypical coverage in patients hospitalised due to CAP. In doing so, this review intends to estimate the mortality and efficacy of regimens containing atypical antibiotic coverage.
Secondary objectives are:
-
To evaluate the frequency of adverse effects associated with different types of antibiotic treatment.
-
To evaluate the efficacy of regimens containing atypical antibiotic coverage for patients with identified atypical pathogens (see later for definition);
-
To evaluate the efficacy of regimens containing atypical antibiotic coverage for patients with identified Streptococcus pneumoniae.
Outcomes will be collected for all patients, and separately for patients with identified atypical pathogens, and those with pneumococcal infection.
Methods
Criteria for considering studies for this review
Types of studies
Any randomised or quasi‐randomised controlled trial, assessing empirical antibiotic treatment of CAP in hospitalised adults. Trials will be included irrespective of publication status, language, and blinding. Trials with a drop out rate of over 30% will be excluded.
Types of participants
Adult patients hospitalised due to suspected CAP. Patients with major immunosuppressive state/s (i.e. malignancy, cystic fibrosis, HIV) will not be considered for inclusion on this review.
Types of interventions
Studies comparing antibiotic therapy with atypical coverage to antibiotic therapy without atypical coverage.
Regimens containing macrolide, fluoroquinolone, tetracycline, doxycycline or chloramphenicol, will be considered atypical coverage. Regimen lacking all of the above, will be considered non‐atypical coverage.
Both oral and intravenous (IV) therapies will be reviewed (whether comparing oral to oral, IV to oral or IV to IV).
Types of outcome measures
Primary outcome measures:
-
Overall mortality at the end of the study and up to 30 days following end of treatment.
Secondary outcome measures:
-
Number of patients with treatment failure, as defined in the study.
-
Number of patients in whom bacteriological eradication was achieved.
-
Number of patients who needed mechanical ventilation during hospital stay.
-
Mean duration of hospital stay.
-
Number of patients excluded from outcome assessment after randomisation
-
Number of patients who developed superinfection or resistance to the studied drug, defined as isolation of a pathogen or a colonising micro‐organism resistant to the study drug, during or after antibiotic treatment.
-
Mean duration to regaining previous functional capacity.
Adverse events:
-
Any serious adverse events that were fatal, life‐threatening, or requiring prolongation of existing hospitalisation.
-
Any adverse events that required discontinuation of medication.
Search methods for identification of studies
See: Collaborative Review Group search strategy
Electronic searching
(i) Relevant randomised clinical trials will be identified by searching the latest issue of the Cochrane Central Register of Controlled Trials (CENTRAL), which includes the Acute Respiratory Infection Groups' specialised register; MEDLINE (January 1966 to present); EMBASE (1980 to present) with the search terms:
ANTI‐INFECTIVE‐AGENTS‐QUINOLONE*:ME or
QUINOLONES*:ME or
FLUOROQUINOLON* or
MACROLIDE* or
ANTIBIOTICS‐MACROLIDE*:ME or
TETRACYCLINE or
DOXYCYCLINE or
CHLORAMPHENICOL or
ANTIBIOTICS‐TETRACYCLINE*:ME or
TETRACYCLINES*:ME AND
COMMUNITY‐ACQUIRED‐INFECTIONS*:ME or
COMMUNIT* and PNEUMONI* or
PNEUMONIA*:ME
The references of all identified studies will be inspected for more trials. Additionally, the first or corresponding author of each included trial and the researchers active in the field will be contacted for information regarding unpublished trials or complementary information on their own trial. In addition to this, major relevant pharmaceutical companies will be contacted for relevant unpublished trials.
Data collection and analysis
Study selection
Two reviewers will independently inspect each reference identified by the search and apply the inclusion criteria. For possible relevant articles, or in cases of disagreement between the two reviewers, the full article will be obtained and inspected independently by the two reviewers. Where resolving disagreement by discussion will not be possible, a third reviewer will check the article. If required, the authors of the study will be contacted for clarification. Justification for excluding studies from the review will also be documented.
Quality assessment
Trials fulfilling the review inclusion criteria will be assessed for methodological quality by two reviewers. Full information will be extracted regarding randomisation and allocation concealment, blinding, sample size, exclusions after randomisation and different lengths of follow‐up. Bias is assessed by allocation concealment (A. Low risk of bias ‐ adequate allocation concealment; B. Moderate risk of bias ‐ unclear allocation concealment; C. High risk of bias ‐ inadequate allocation concealment, i.e., quasi‐randomised studies). Randomization is graded A through C (A ‐ table of random numbers, computer generated or coin tossing; B‐unclear; C‐anything else); Blinding is classified A‐D (A ‐ triple blinded, B ‐ double blinded, C ‐ patient blinded, D ‐ no blinding).
Sensitivity analyses will be performed in order to assess the robustness of the findings to different aspects of the trials methodology: allocation concealment (adequate or unclear), randomisation, blinding, exclusions after randomisation (reported or not reported), sample size (up to 100 or more than 100 patients) and length of follow up (up to one month or one to 12 months).
Data collection
Two reviewers will independently extract the data of included trials. In case of any disagreement between the two reviewers, a third reviewer will extract the data. The data extraction will be discussed, decisions documented and, where necessary, the authors of the trials will be contacted for clarification.
Trials will be identified by the name of the first author and year in which the trial was first published and ordered chronologically.
The following data will be extracted, checked, and recorded:
Trial characteristics:
1. Year/s and country/s of study
2. Trial sponsor
3. Publication status: published in journal; abstract/ proceeding; unpublished
4. Intention to treat analysis: performed; possible to extract; efficacy analysis
5. Randomisation methods: allocation generation; concealment; blinding
6. Failure definition: including time of failure assessment
7. Study follow‐up duration
Patient characteristics:
8. Number of patients randomised and evaluated, per group
9. Age: mean or median
Infection characteristics:
10. Number of patients with documented typical pathogen
11. Number of patients with documented atypical pathogen
12. Number of patients with infections caused by bacteriae resistant to the administered antibiotic regimen
13. Patients with pneumococcus , legionella and mycoplasma infections
14. Severity of the disease (as classified by PORT)
Intervention characteristics:
15. Antibiotics type, dose and route of administration
16. Treatment duration
17. Treatment modifications
Measures of outcome, extracted as number of patients per group:
18. Overall mortality at end of study follow‐up
19. Bacteriological eradication
20. Treatment failure ‐ as defined in study, with and without treatment modifications
21. Mechanical ventilation
22. Adverse events: life threatening events, events necessitating specific medical treatment, events requiring cessation of treatment.
Data synthesis
Dichotomous data will be analysed by calculating the relative risk (RR) for each trial with the uncertainty in each result being expressed using 95% confidence intervals (CI).
Continuous data will be analysed by using the mean and standard deviation of each trial and calculating the effect size (average mean difference) and the 95% CI, whenever comparisons made between the mean duration of symptoms in the two groups are normally distributed.
Heterogeneity and publication bias
Heterogeneity in the results of the trials will initially be assessed by inspection of graphical presentations and by calculating a test of heterogeneity (Chi‐square). We anticipate between‐trial variation in estimation of morbidity and mortality for those patients who were hospitalised on a general ward or in intensive care units; if trials were performed in different geographical areas, and among different age groups; per different drug regimens assessed. Subgroup analyses will be performed in order to assess the impact of these possible sources of heterogeneity on the main results.
A fixed effect model will be used throughout the review, except in the event of significant heterogeneity between the trials (P < 0.10), when the random effect model will be chosen. In addition, a funnel plot estimating the precision of trials (plots of logarithm of the relative risk (RR) for efficacy against the sample size) will be examined in order to estimate potential asymmetry.
